Last updated: November 2, 2026
Szymon Niemiec operates Transcription: Voice Recorder (the “App”) for iPhone and iPad. For privacy questions or requests, email [email protected]. We are the controller of personal data we receive for support and the operation of the App, where applicable. You do not need an App account to record or transcribe.
Audio transcription takes place on your device. This does not mean the App never uses the internet: model downloads, purchase verification, usage analytics, support email and these web pages involve separate services described below.
The App processes microphone recordings, imported audio and supported video, transcripts, speaker labels, edits, search queries, folders, preferences and optional AI-generated summaries locally. The required models must be downloaded first. We do not upload your recordings or transcripts to a cloud transcription service, receive them for AI training, or provide an App cloud-sync service.
Speaker analysis groups and labels voices within a recording. It is not an identity-verification service and does not establish a speaker's real-world identity. Names you assign to speakers remain part of your local notes. Summaries and other generated output can contain errors; review important content against the source.
The App stores content, job progress, preferences and downloaded models on your device. Files selected from Files or another provider may be copied into the App's local storage. The provider you select may separately use cloud storage. Operating-system backups, device transfers, files you export and copies you make can retain content outside the App. Manage these using the relevant device, account or provider settings.
Microphone access is used when you choose to record. You can change permission in system Settings; importing a supported file does not require recording through the microphone. Notifications are optional and used for local processing-status notices. The App does not require contact-list or location access for transcription.
When you choose to export, copy or share content, the App prepares the selected data locally and you select its destination. Clipboard contents, exported archives, audio, TXT, Markdown and SRT files can contain personal or confidential information. The destination app, person or service handles the copy under its own practices. We cannot delete copies you shared elsewhere.
The App downloads speech, speaker-analysis and summary models from Hugging Face URLs and stores them locally. Download hosts and their delivery/network providers can receive your IP address and ordinary connection information, such as request time, requested file and client information. Your recordings and transcript text are not included in model download requests. Removing a model may require downloading it again before its feature can be used.
Apple processes App Store payments. RevenueCat, when configured for the App, helps display offers, validate purchases, restore access and manage paid entitlements. It can process a generated customer identifier, app/platform information, product identifiers, transaction or receipt information, subscription status and expiry. Purchase verification can require a connection and can occur when offers or entitlements are checked, including before a purchase. We do not receive your full payment-card or bank-account details.
Purchase information is used to provide and restore paid features, administer purchases, understand purchase/subscription performance and prevent fraud. This is separate from local audio processing. Where the GDPR or UK GDPR applies, requested purchase and entitlement processing generally relies on the contract or requested pre-contract steps; proportionate fraud prevention and purchase administration can rely on legitimate interests, and legally required records on legal obligations.
Basic product analytics is enabled automatically when the App starts, with no in-app analytics switch. PostHog EU Cloud receives manually defined app opens, fixed screen categories, welcome-guide progress, model-setup results, recording/import actions, processing results, exports, purchase actions and restore results. Events include a randomly generated installation identifier, session identifier, event time, app version/build and operating-system version. These identifiers allow activity from the same installation to be grouped; they are not your name, email, Apple Account, advertising identifier or RevenueCat customer identifier. We do not connect the PostHog identifier to purchase-account identities.
Audio, transcripts, summaries, note titles, speaker names, custom terms, search text, file names, paths and note identifiers are never included in analytics. Automatic tap/screen capture, session replay, automatic error/crash uploads, surveys and person profiles are disabled. Processing failures are recorded only as fixed event categories, without error messages or content. GeoIP enrichment is disabled and the analytics payload substitutes a zero IP address; PostHog and its network providers still necessarily receive the source IP when delivering HTTPS requests.
Analytics is used to understand feature adoption, setup and processing reliability, and purchase-flow outcomes so we can improve the App. Collection starts automatically after the analytics identifier has been initialized; there is no in-app opt-in or opt-out control. The random analytics identifier is shown in Settings > Help & information > Analytics details. Include this identifier when contacting support about analytics, requesting deletion or exercising applicable privacy rights, and keep a copy before uninstalling. Events already received are not automatically erased by uninstalling. Analytics records are retained according to the PostHog project's configured retention and can be removed in response to an identifiable request, subject to applicable legal duties. Contact us for the current retention period.